Message ID | pull.1299.git.git.1658855372189.gitgitgadget@gmail.com (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | config.c: NULL check when reading protected config | expand |
On Tue, Jul 26, 2022 at 05:09:32PM +0000, Glen Choo via GitGitGadget wrote: > From: Glen Choo <chooglen@google.com> > > In read_protected_config(), check whether each file name is NULL before > attempting to read it. This mirrors do_git_config_sequence() (which > read_protected_config() is modelled after). s/modelled/modeled > Without these NULL checks, > > make SANITIZE=address test T=t0410*.sh I'm glad that t0410 was catching this for us already, though it is too bad we didn't see it outside of the ASan builds, or I think we could have potentially caught this earlier. Either way, I think the test coverage here is sufficient, so what you wrote makes sense. > diff --git a/config.c b/config.c > index 015bec360f5..b0ba7f439a4 100644 > --- a/config.c > +++ b/config.c > @@ -2645,9 +2645,12 @@ static void read_protected_config(void) > system_config = git_system_config(); > git_global_config(&user_config, &xdg_config); > > - git_configset_add_file(&protected_config, system_config); > - git_configset_add_file(&protected_config, xdg_config); > - git_configset_add_file(&protected_config, user_config); > + if (system_config) > + git_configset_add_file(&protected_config, system_config); > + if (xdg_config) > + git_configset_add_file(&protected_config, xdg_config); > + if (user_config) > + git_configset_add_file(&protected_config, user_config); > git_configset_add_parameters(&protected_config); I wonder: should it become a BUG() to call git_configset_add_file() with a NULL filename? That would have elevated the test failure outside of just the ASAn builds, I'd think. There's certainty a risk of being too defensive, but elevating this error beyond just the ASan builds indicates that this would be an appropriate layer of defense IMHO. Thanks, Taylor
Taylor Blau <me@ttaylorr.com> writes: > On Tue, Jul 26, 2022 at 05:09:32PM +0000, Glen Choo via GitGitGadget wrote: >> From: Glen Choo <chooglen@google.com> >> >> In read_protected_config(), check whether each file name is NULL before >> attempting to read it. This mirrors do_git_config_sequence() (which >> read_protected_config() is modelled after). > > s/modelled/modeled Ah, thanks. >> Without these NULL checks, >> >> make SANITIZE=address test T=t0410*.sh > > I'm glad that t0410 was catching this for us already, though it is too > bad we didn't see it outside of the ASan builds, or I think we could > have potentially caught this earlier. > > Either way, I think the test coverage here is sufficient, so what you > wrote makes sense. > >> diff --git a/config.c b/config.c >> index 015bec360f5..b0ba7f439a4 100644 >> --- a/config.c >> +++ b/config.c >> @@ -2645,9 +2645,12 @@ static void read_protected_config(void) >> system_config = git_system_config(); >> git_global_config(&user_config, &xdg_config); >> >> - git_configset_add_file(&protected_config, system_config); >> - git_configset_add_file(&protected_config, xdg_config); >> - git_configset_add_file(&protected_config, user_config); >> + if (system_config) >> + git_configset_add_file(&protected_config, system_config); >> + if (xdg_config) >> + git_configset_add_file(&protected_config, xdg_config); >> + if (user_config) >> + git_configset_add_file(&protected_config, user_config); >> git_configset_add_parameters(&protected_config); > > I wonder: should it become a BUG() to call git_configset_add_file() with > a NULL filename? That would have elevated the test failure outside of > just the ASAn builds, I'd think. > > There's certainty a risk of being too defensive, but elevating this > error beyond just the ASan builds indicates that this would be an > appropriate layer of defense IMHO. Hm, if we're going in this direction, what if we made it a BUG() to call fopen_or_warn() with a NULL filename? Then we wouldn't have to reimplement this BUG() check in all of its callers. > > Thanks, > Taylor
On Tue, Jul 26, 2022 at 10:40:18AM -0700, Glen Choo wrote: > > I wonder: should it become a BUG() to call git_configset_add_file() with > > a NULL filename? That would have elevated the test failure outside of > > just the ASAn builds, I'd think. > > > > There's certainty a risk of being too defensive, but elevating this > > error beyond just the ASan builds indicates that this would be an > > appropriate layer of defense IMHO. > > Hm, if we're going in this direction, what if we made it a BUG() to call > fopen_or_warn() with a NULL filename? Then we wouldn't have to > reimplement this BUG() check in all of its callers. That may be too low-level of a place to put this check, but I don't have a strong opinion about it either way (including whether we should have such a BUG() *anywhere* in this series, including git_configset_add_file()). Thanks, Taylor
On 7/26/2022 1:43 PM, Taylor Blau wrote: > On Tue, Jul 26, 2022 at 10:40:18AM -0700, Glen Choo wrote: >>> I wonder: should it become a BUG() to call git_configset_add_file() with >>> a NULL filename? That would have elevated the test failure outside of >>> just the ASAn builds, I'd think. >>> >>> There's certainty a risk of being too defensive, but elevating this >>> error beyond just the ASan builds indicates that this would be an >>> appropriate layer of defense IMHO. >> >> Hm, if we're going in this direction, what if we made it a BUG() to call >> fopen_or_warn() with a NULL filename? Then we wouldn't have to >> reimplement this BUG() check in all of its callers. > > That may be too low-level of a place to put this check, but I don't have > a strong opinion about it either way (including whether we should have > such a BUG() *anywhere* in this series, including > git_configset_add_file()). Since git_configset_add_file() returns an 'int', could we return -1 if the supplied 'filename' was null? (The correct place to check would be down in git_config_from_file_with_options().) It would save all these checks here. (Also: do we care that we are ignoring the return values in read_protected_config()? Thanks, -Stolee
On Tue, Jul 26 2022, Glen Choo via GitGitGadget wrote: > From: Glen Choo <chooglen@google.com> > > In read_protected_config(), check whether each file name is NULL before > attempting to read it. This mirrors do_git_config_sequence() (which > read_protected_config() is modelled after). > > Without these NULL checks, > > make SANITIZE=address test T=t0410*.sh > > fails because xdg_config is NULL, causing us to call fopen(NULL). FWIW a lot more than that fails, that's just the test I focused on for the bug report, the others ones (I didn't check out all of them) all variants of that. See https://github.com/avar/git/runs/7519070124?check_suite_focus=true for the current failing run with that "[2]" patch you quoted. We fail a total of 14 test files (and many more tests within those files). > Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> > Signed-off-by: Glen Choo <chooglen@google.com> > --- > config.c: NULL check when reading protected config > > This fixes the SANITIZE=address failure on master, That was introduced > by gc/bare-repo-discovery. Thanks again to Ævar for the original report > [1] and for proposing a way to catch this in CI [2]. > > [1] > https://lore.kernel.org/git/220725.861qu9oxl4.gmgdl@evledraar.gmail.com > [2] > https://lore.kernel.org/git/patch-1.1-e48b6853dd5-20220726T110716Z-avarab@gmail.com > > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1299%2Fchooglen%2Fconfig%2Ffix-sanitize-address-v1 > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1299/chooglen/config/fix-sanitize-address-v1 > Pull-Request: https://github.com/git/git/pull/1299 > > config.c | 9 ++++++--- > 1 file changed, 6 insertions(+), 3 deletions(-) > > diff --git a/config.c b/config.c > index 015bec360f5..b0ba7f439a4 100644 > --- a/config.c > +++ b/config.c > @@ -2645,9 +2645,12 @@ static void read_protected_config(void) > system_config = git_system_config(); > git_global_config(&user_config, &xdg_config); > > - git_configset_add_file(&protected_config, system_config); > - git_configset_add_file(&protected_config, xdg_config); > - git_configset_add_file(&protected_config, user_config); > + if (system_config) > + git_configset_add_file(&protected_config, system_config); > + if (xdg_config) > + git_configset_add_file(&protected_config, xdg_config); > + if (user_config) > + git_configset_add_file(&protected_config, user_config); > git_configset_add_parameters(&protected_config); > > free(system_config); > > base-commit: 6a475b71f8c4ce708d69fdc9317aefbde3769e25 Re your claim in https://lore.kernel.org/git/kl6lzggwsyh1.fsf@chooglen-macbookpro.roam.corp.google.com/ I tried testing this, and came up with the below. I wonder if we should work in here for general paranoia, but I'm not too familiar with the this part of the config reading, maybe we're confident enough that these are invariants within the process. This will BUG() out if these variables change within the process, which would mean that our caching assumptions are no longer true, which would cause you to return the wrong data here. Of course you'd have segfaulted or similar before, but this should demonstrate that not only are these sometimes NULL, but that they stay that way. diff --git a/config.c b/config.c index 015bec360f5..cdd665c1cc8 100644 --- a/config.c +++ b/config.c @@ -2102,6 +2102,30 @@ int git_config_system(void) return !git_env_bool("GIT_CONFIG_NOSYSTEM", 0); } +static char *last_system_config; +static char *last_xdg_config; +static char *last_user_config; + +static void sanity_check_config_1(const char *name, char **last, + const char *now) +{ + if (*last && now && strcmp(*last, now)) + BUG("%s_config: had '%s', now '%s'", name, *last, now); + else if (*last && !now) + BUG("%s_config: had '%s', now NULL", name, *last); + free(*last); + *last = xstrdup_or_null(now); +} + +static void sanity_check_config(const char *system_config, + const char *xdg_config, + const char *user_config) +{ + sanity_check_config_1("system", &last_system_config, system_config); + sanity_check_config_1("xdg", &last_xdg_config, xdg_config); + sanity_check_config_1("user", &last_user_config, user_config); +} + static int do_git_config_sequence(const struct config_options *opts, config_fn_t fn, void *data) { @@ -2134,6 +2158,8 @@ static int do_git_config_sequence(const struct config_options *opts, if (user_config && !access_or_die(user_config, R_OK, ACCESS_EACCES_OK)) ret += git_config_from_file(fn, user_config, data); + sanity_check_config(system_config, xdg_config, user_config); + current_parsing_scope = CONFIG_SCOPE_LOCAL; if (!opts->ignore_repo && repo_config && !access_or_die(repo_config, R_OK, 0)) @@ -2645,6 +2671,8 @@ static void read_protected_config(void) system_config = git_system_config(); git_global_config(&user_config, &xdg_config); + sanity_check_config(system_config, xdg_config, user_config); + git_configset_add_file(&protected_config, system_config); git_configset_add_file(&protected_config, xdg_config); git_configset_add_file(&protected_config, user_config);
Derrick Stolee <derrickstolee@github.com> writes: > On 7/26/2022 1:43 PM, Taylor Blau wrote: >> On Tue, Jul 26, 2022 at 10:40:18AM -0700, Glen Choo wrote: >>>> I wonder: should it become a BUG() to call git_configset_add_file() with >>>> a NULL filename? That would have elevated the test failure outside of >>>> just the ASAn builds, I'd think. >>>> >>>> There's certainty a risk of being too defensive, but elevating this >>>> error beyond just the ASan builds indicates that this would be an >>>> appropriate layer of defense IMHO. >>> >>> Hm, if we're going in this direction, what if we made it a BUG() to call >>> fopen_or_warn() with a NULL filename? Then we wouldn't have to >>> reimplement this BUG() check in all of its callers. >> >> That may be too low-level of a place to put this check, but I don't have >> a strong opinion about it either way (including whether we should have >> such a BUG() *anywhere* in this series, including >> git_configset_add_file()). > > Since git_configset_add_file() returns an 'int', could we return -1 > if the supplied 'filename' was null? (The correct place to check would > be down in git_config_from_file_with_options().) > > It would save all these checks here. Hm, IIUC you are suggesting that git_configset_add_file() returns -1 instead of BUG()-ing? BUG() sounds better IMO, since there really is nothing useful that git_configset_add_file() (and later functions) can do with a NULL file name. Plus, git_configset_add_file() has already reserved -1 to mean "a file was specified but could not be read". > > (Also: do we care that we are ignoring the return values in > read_protected_config()? I don't think we care (unless this is a style issue). "git config" succeeds even if it encounters non-repo files that can't be read. In a similar vein, I don't think it matters for protected config if we can't read one of the files (e.g. xdg_config) or even all of the files; all that matters is that we've read everything that we can. > > Thanks, > -Stolee
Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes: > On Tue, Jul 26 2022, Glen Choo via GitGitGadget wrote: > >> From: Glen Choo <chooglen@google.com> >> >> In read_protected_config(), check whether each file name is NULL before >> attempting to read it. This mirrors do_git_config_sequence() (which >> read_protected_config() is modelled after). >> >> Without these NULL checks, >> >> make SANITIZE=address test T=t0410*.sh >> >> fails because xdg_config is NULL, causing us to call fopen(NULL). > > FWIW a lot more than that fails, that's just the test I focused on for > the bug report, the others ones (I didn't check out all of them) all > variants of that. > > See https://github.com/avar/git/runs/7519070124?check_suite_focus=true > for the current failing run with that "[2]" patch you quoted. We fail a > total of 14 test files (and many more tests within those files). Ah thanks, I'll amend the message accordingly. >> Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> >> Signed-off-by: Glen Choo <chooglen@google.com> >> --- >> config.c: NULL check when reading protected config >> >> This fixes the SANITIZE=address failure on master, That was introduced >> by gc/bare-repo-discovery. Thanks again to Ævar for the original report >> [1] and for proposing a way to catch this in CI [2]. >> >> [1] >> https://lore.kernel.org/git/220725.861qu9oxl4.gmgdl@evledraar.gmail.com >> [2] >> https://lore.kernel.org/git/patch-1.1-e48b6853dd5-20220726T110716Z-avarab@gmail.com >> >> Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1299%2Fchooglen%2Fconfig%2Ffix-sanitize-address-v1 >> Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1299/chooglen/config/fix-sanitize-address-v1 >> Pull-Request: https://github.com/git/git/pull/1299 >> >> config.c | 9 ++++++--- >> 1 file changed, 6 insertions(+), 3 deletions(-) >> >> diff --git a/config.c b/config.c >> index 015bec360f5..b0ba7f439a4 100644 >> --- a/config.c >> +++ b/config.c >> @@ -2645,9 +2645,12 @@ static void read_protected_config(void) >> system_config = git_system_config(); >> git_global_config(&user_config, &xdg_config); >> >> - git_configset_add_file(&protected_config, system_config); >> - git_configset_add_file(&protected_config, xdg_config); >> - git_configset_add_file(&protected_config, user_config); >> + if (system_config) >> + git_configset_add_file(&protected_config, system_config); >> + if (xdg_config) >> + git_configset_add_file(&protected_config, xdg_config); >> + if (user_config) >> + git_configset_add_file(&protected_config, user_config); >> git_configset_add_parameters(&protected_config); >> >> free(system_config); >> >> base-commit: 6a475b71f8c4ce708d69fdc9317aefbde3769e25 > > Re your claim in > https://lore.kernel.org/git/kl6lzggwsyh1.fsf@chooglen-macbookpro.roam.corp.google.com/ > I tried testing this, and came up with the below. > > I wonder if we should work in here for general paranoia, but I'm not too > familiar with the this part of the config reading, maybe we're confident > enough that these are invariants within the process. > > This will BUG() out if these variables change within the process, which > would mean that our caching assumptions are no longer true, which would > cause you to return the wrong data here. > > Of course you'd have segfaulted or similar before, but this should > demonstrate that not only are these sometimes NULL, but that they stay > that way. Interesting, this is worth proposing, but I suspect that the conversation will be long enough for this to be its own thread. Surely someone must have given some thought to this, especially for long-lived processes (git-daemon?). There's also the general question of config cache freshness, e.g. what if another git process writes to a shared config file? (We don't worry about the single process case because the process will invalidate its own cache). Perhaps we should also worry about that (probably more common) case in addition to this one? At any rate, that seems like a bigger topic than this fix here. > > diff --git a/config.c b/config.c > index 015bec360f5..cdd665c1cc8 100644 > --- a/config.c > +++ b/config.c > @@ -2102,6 +2102,30 @@ int git_config_system(void) > return !git_env_bool("GIT_CONFIG_NOSYSTEM", 0); > } > > +static char *last_system_config; > +static char *last_xdg_config; > +static char *last_user_config; > + > +static void sanity_check_config_1(const char *name, char **last, > + const char *now) > +{ > + if (*last && now && strcmp(*last, now)) > + BUG("%s_config: had '%s', now '%s'", name, *last, now); > + else if (*last && !now) > + BUG("%s_config: had '%s', now NULL", name, *last); > + free(*last); > + *last = xstrdup_or_null(now); > +} > + > +static void sanity_check_config(const char *system_config, > + const char *xdg_config, > + const char *user_config) > +{ > + sanity_check_config_1("system", &last_system_config, system_config); > + sanity_check_config_1("xdg", &last_xdg_config, xdg_config); > + sanity_check_config_1("user", &last_user_config, user_config); > +} > + > static int do_git_config_sequence(const struct config_options *opts, > config_fn_t fn, void *data) > { > @@ -2134,6 +2158,8 @@ static int do_git_config_sequence(const struct config_options *opts, > if (user_config && !access_or_die(user_config, R_OK, ACCESS_EACCES_OK)) > ret += git_config_from_file(fn, user_config, data); > > + sanity_check_config(system_config, xdg_config, user_config); > + > current_parsing_scope = CONFIG_SCOPE_LOCAL; > if (!opts->ignore_repo && repo_config && > !access_or_die(repo_config, R_OK, 0)) > @@ -2645,6 +2671,8 @@ static void read_protected_config(void) > system_config = git_system_config(); > git_global_config(&user_config, &xdg_config); > > + sanity_check_config(system_config, xdg_config, user_config); > + > git_configset_add_file(&protected_config, system_config); > git_configset_add_file(&protected_config, xdg_config); > git_configset_add_file(&protected_config, user_config);
On Tue, Jul 26 2022, Glen Choo wrote: > Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes: > >> On Tue, Jul 26 2022, Glen Choo via GitGitGadget wrote: >> >>> From: Glen Choo <chooglen@google.com> >>> >>> In read_protected_config(), check whether each file name is NULL before >>> attempting to read it. This mirrors do_git_config_sequence() (which >>> read_protected_config() is modelled after). >>> >>> Without these NULL checks, >>> >>> make SANITIZE=address test T=t0410*.sh >>> >>> fails because xdg_config is NULL, causing us to call fopen(NULL). >> >> FWIW a lot more than that fails, that's just the test I focused on for >> the bug report, the others ones (I didn't check out all of them) all >> variants of that. >> >> See https://github.com/avar/git/runs/7519070124?check_suite_focus=true >> for the current failing run with that "[2]" patch you quoted. We fail a >> total of 14 test files (and many more tests within those files). > > Ah thanks, I'll amend the message accordingly. > >>> Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> >>> Signed-off-by: Glen Choo <chooglen@google.com> >>> --- >>> config.c: NULL check when reading protected config >>> >>> This fixes the SANITIZE=address failure on master, That was introduced >>> by gc/bare-repo-discovery. Thanks again to Ævar for the original report >>> [1] and for proposing a way to catch this in CI [2]. >>> >>> [1] >>> https://lore.kernel.org/git/220725.861qu9oxl4.gmgdl@evledraar.gmail.com >>> [2] >>> https://lore.kernel.org/git/patch-1.1-e48b6853dd5-20220726T110716Z-avarab@gmail.com >>> >>> Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1299%2Fchooglen%2Fconfig%2Ffix-sanitize-address-v1 >>> Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1299/chooglen/config/fix-sanitize-address-v1 >>> Pull-Request: https://github.com/git/git/pull/1299 >>> >>> config.c | 9 ++++++--- >>> 1 file changed, 6 insertions(+), 3 deletions(-) >>> >>> diff --git a/config.c b/config.c >>> index 015bec360f5..b0ba7f439a4 100644 >>> --- a/config.c >>> +++ b/config.c >>> @@ -2645,9 +2645,12 @@ static void read_protected_config(void) >>> system_config = git_system_config(); >>> git_global_config(&user_config, &xdg_config); >>> >>> - git_configset_add_file(&protected_config, system_config); >>> - git_configset_add_file(&protected_config, xdg_config); >>> - git_configset_add_file(&protected_config, user_config); >>> + if (system_config) >>> + git_configset_add_file(&protected_config, system_config); >>> + if (xdg_config) >>> + git_configset_add_file(&protected_config, xdg_config); >>> + if (user_config) >>> + git_configset_add_file(&protected_config, user_config); >>> git_configset_add_parameters(&protected_config); >>> >>> free(system_config); >>> >>> base-commit: 6a475b71f8c4ce708d69fdc9317aefbde3769e25 >> >> Re your claim in >> https://lore.kernel.org/git/kl6lzggwsyh1.fsf@chooglen-macbookpro.roam.corp.google.com/ >> I tried testing this, and came up with the below. >> >> I wonder if we should work in here for general paranoia, but I'm not too >> familiar with the this part of the config reading, maybe we're confident >> enough that these are invariants within the process. >> >> This will BUG() out if these variables change within the process, which >> would mean that our caching assumptions are no longer true, which would >> cause you to return the wrong data here. >> >> Of course you'd have segfaulted or similar before, but this should >> demonstrate that not only are these sometimes NULL, but that they stay >> that way. > > Interesting, this is worth proposing, but I suspect that the > conversation will be long enough for this to be its own thread. Surely > someone must have given some thought to this, especially for long-lived > processes (git-daemon?). > > There's also the general question of config cache freshness, e.g. what > if another git process writes to a shared config file? (We don't worry > about the single process case because the process will invalidate its > own cache). > > Perhaps we should also worry about that (probably more common) case in > addition to this one? At any rate, that seems like a bigger topic than > this fix here. We can leave it for later, I've run it as a one-off and didn't have any failures. But FWIW I think it's tied up in this fix here, i.e. your original code both added caching, and implicitly assumed that these were never NULL, so it was "obvious" that it didn't need such assertions. But now we have 3x if's in a code path that's cached, and the cache is *not* guarded by the same 3x checks. So we can leave it for later, but it really seems worth adding some self-documentation here sooner than later. This BUG() method I came up with is one way, another would be to strdup it and use a "static" variable in the function, i.e. stick with whatever value(s) we start out with. But in any case, this fix seems correct, and fixes the current issues SANITIZE=address is spotting for us, thanks!
diff --git a/config.c b/config.c index 015bec360f5..b0ba7f439a4 100644 --- a/config.c +++ b/config.c @@ -2645,9 +2645,12 @@ static void read_protected_config(void) system_config = git_system_config(); git_global_config(&user_config, &xdg_config); - git_configset_add_file(&protected_config, system_config); - git_configset_add_file(&protected_config, xdg_config); - git_configset_add_file(&protected_config, user_config); + if (system_config) + git_configset_add_file(&protected_config, system_config); + if (xdg_config) + git_configset_add_file(&protected_config, xdg_config); + if (user_config) + git_configset_add_file(&protected_config, user_config); git_configset_add_parameters(&protected_config); free(system_config);